Restless legs syndrome (RLS), recently renamed Willis-Ekbom disease (WED), is a sensorimotor disorder with a primary sensory symptom of a strong, often irresistible urge to move the legs. The operational diagnostic criteria for RLS of the International RLS Study Group (IRLSSG) rely on the presence of four essential clinical features, including 1) an urge to move the legs caused by uncomfortable and unpleasant sensations; 2) the symptoms begin or worsen during rest or inactivity; 3) the symptoms are partially or totally relieved by movements; and 4) the symptoms only occur or are worse in the evening or night than during the day.(1) The symptoms are not solely accounted for as being primary to another condition, such as leg cramps or positional discomfort.(2) In addition to the essential criteria, three supportive factors have been identified: 1) a positive family history of RLS; 2) a positive response to dopaminergic drugs; and 3) periodic limb movements during wakefulness and sleep.(1) Recent epidemiologic surveys of RLS in the general population reported a prevalence of 7.2% for any form of RLS (primary or secondary) and 2.4% for primary RLS, making it the most common movement disorder and among the most common of the sleep disorders.(3, 4)
A large number of disorders have been suggested to be associated with RLS. The evidence for some is stronger than for others, which poses the important question of whether these disorders are comorbid, coincidental, or secondary causes of RLS. Indeed, many of these conditions share an association with depleted iron stores (iron deficiency and anemia), renal failure, pregnancy, rheumatoid arthritis, peripheral neuropathies, Parkinson’s disease, and liver disease, while others cannot yet be explained on the basis of a single pathophysiological mechanism.(5) In addition to the above causes, there are many reports in the literature asserting pharmacologically induced RLS. One of the earliest reports of medication-induced RLS was one of Ekbom’s original descriptions of a survey of his 175 personal cases that RLS can be caused by the intake of certain drugs including promethazine and prochlorperazine.(6) Early studies also suggested that RLS can be induced by certain substances and medications, such as caffeine and beta-blockers.(7, 8) Caffeine was claimed as a cause of RLS due to its propensity to increase nervous system arousal and the direct peripheral contractile effect on the striated muscle.(8) However, it is difficult to determine the true causation of those medications in the early reports since the information on the diagnosis of RLS and the temporal associations were often incomplete.(9, 10) Following the early descriptions, there has continued to be a flow of publications that have identified a variety of medications that can induce or aggravate RLS. Some drugs used to treat RLS, especially dopaminergic medications, can also aggravate the symptoms in the form of what we now know as augmentation. However, the vast majority of the literature on medication-induced RLS are case reports and the quality of available evidence varies. In order to establish true causation for medication-induced RLS, there should be no history of RLS prior to drug initiation, secondary causes should be excluded, dosage timing close to bedtime to explain nocturnal symptoms, endorsement of all four 2003 National Institute of Health (NIH) criteria for definitive diagnosis of RLS, and a polysomnogram (PSG) should be administered to rule out sleep-disordered breathing as a cause of nocturnal disturbance that may be associated with RLS.(11) The most important evidence for etiologic determination comes from trials on and off the offending medication with clinical reassessments for changes in RLS. Despite the large number of studies, very few articles applied the above criteria to determine the true causation of RLS.
Awareness of the medications that can potentially lead to RLS is crucial because it changes treatment strategies. Instead of starting another medication such as a dopamine agonist to treat iatrogenic RLS, it may be more prudent to withdraw the potentially offending medication as the first line intervention. In general, medication-induced RLS should resolve when the dosage is reduced or the drug is withdrawn. A new dose increase, or rechallenge with the same drug should be followed by symptom recurrence.
A basic understanding of the possible mechanisms involved in RLS will help us understand why certain classes of medications are more likely to induce or precipitate RLS than others. Although the exact mechanism of RLS is still unknown, several lines of evidence from neuroimaging, neuropathology, CSF, and genetic linkage studies favor altered brain iron-dopamine mechanisms. This association is supported by human studies of striatal D2 receptors and iron reduction, increased nigral tyrosine hydroxylase, reduced CSF ferritin, and human genome-wide association studies implicating linkages to several susceptible loci. Abnormal iron stores have been correlated with altered central neurotransmitter systems, as dopamine production requires ferritin as a cofactor for tyrosine hydroxylase, the rate-limiting enzyme in dopamine production.(12) Furthermore, recent evidence also suggests that there is a probable association of peripheral tissue hypoxia and dopaminergic neurons outside of the blood-brain barrier and RLS.(13, 14) The above evidence supports the understanding that RLS is a network disorder that involves many regions of the nervous system, from the periphery to the cortex, and contains structures that are involved in somatosensory perception as well as the generation of movement.(12)
Since the impressive efficacy of dopaminergic drugs for the treatment of RLS supports the dysfunction of the dopaminergic system in the pathophysiology of RLS, it is reasonable to assume that dopamine receptor blocking agents (DRBAs), which are primarily antipsychotics, can induce RLS. On the other hand, the mechanism of antidepressant-induced RLS is more difficult to explain since most antidepressants process α1-adrenoreceptor blockade (mainly the tricyclic antidepressants), α2-adrenoreceptor blockade (mirtazapine), and very weak D2-receptor antagonistic activity. Some antidepressants also exhibit histamine H1, muscarinic, and 5-HT2A receptor blockades. Therefore, none of these actions would be expected to provoke RLS through the dopaminergic system directly. Drugs that selectively inhibit serotonin transporters have been shown to downregulate dopaminergic and norepinephric neurons and this mechanism may be partly responsible for the development of extrapyramidal syndromes associated with the use of these medications.(15) Nevertheless, the exact mechanism in which each class of medications contributes to RLS is largely unknown and it is likely that more than one mechanism is involved which facilitates abnormal dopamine/serotonin interactions. In addition, individual susceptibility, as well as pharmacologic profiles of individual medications, needs to be considered as not all patients who are exposed to these medications will develop RLS. Detailed discussions are provided below on each class of drugs.
Among various medications that have been shown to cause RLS, two separate systematic reviews have identified similar groups of medications that have been frequently associated with RLS. The first study evaluated the evidence with the least amount of confounding by rating the quality of individual articles according to strict inclusion and exclusion criteria. It identified strong evidence for medication-induced RLS for the following drugs: escitalopram; fluoxetine; L-dopa/carbidopa and pergolide; L-thyroxine; mianserin; mirtazapine; olanzapine; and tramadol.(11) Another systematic review identified medications that induced RLS as an adverse reaction from the French pharmacovigilance database. The most reported drugs were antidepressants (amitriptyline, escitalopram, mianserin, mirtazapine, duloxetine), antipsychotics (thioridazine, loxapine, risperidone, and aripiprazole), and tramadol.(16) The common occurrence of these medications to induce RLS may give us some insights into the pathophysiological mechanism of RLS. For the purpose of the review in this chapter, we have categorized the medications that can cause RLS into five main groups: 1) antidepressants, 2) antipsychotics, 3) antiepileptic drugs, 4) opioids and opioid receptor agonists, and 5) miscellaneous (Table 9.1). In this chapter, we also review the issue of augmentation, which is the progressive exacerbation of RLS earlier in the day after administration of dopaminergic medication in the afternoon or evening.(17)
2) Dopamine receptor blocking agents
Antidepressant medications constitute different classes of medications but the frequently used groups in clinical practice are tricyclic antidepressants (TCAs), selective serotonin reuptake inhibitors (SSRIs), serotonin-norepinephrine reuptake inhibitors (SNRIs), alpha 2 antagonists as serotonin and norepinephrine disinhibitors (SNDIs), and monoamine oxidase inhibitors (MAOIs). The principle mechanism of all effective antidepressants involves boosting the synaptic action of one or more of the monoamines, mainly dopamine, serotonin, and norepinephrine. This action is often but not exclusively conducted by acutely blocking one or more of the presynaptic transporters of these monoamines.(18) MAOIs were the first class of antidepressants to be developed, but fell out of favor because of concerns about interactions with certain foods and numerous drug interactions. RLS has been reported as a possible side effect in a number of TCAs (amitriptyline), SSRIs (fluoxetine, paroxetine, sertraline, citalopram escitalopram), SNRIs (venlafaxine, duloxetine), and others (mianserin, mirtazapine), but the majority of evidence is in the form of case reports.(19–32) Among those reports, frequent associations were observed on mirtazapine and mianserin, which share similar 5-HT2 antagonistic activity, supporting the possibility that the different effects of noradrenergic and serotonergic transmission are involved in the development of RLS.(26, 28, 30–32) One retrospective chart review indicated that the risk of developing RLS among mirtazapine users was approximately 8% and the concomitant use of medication with tramadol, fluoxetine, or dopamine receptor blocking agents (DRBAs, including domperidone which is a peripheral dopamine D2 receptor antagonist) could potentially increase the risk.(26, 27, 33) To the best of our knowledge, no cases of RLS have been reported as an adverse event of reboxetine or bupropion.
Despite a number of case reports suggesting that some antidepressants may induce RLS, a few systematic studies reported contradictory results. In a retrospective study looking at the association between the clinical diagnosis of RLS and the use of antidepressants in 200 patients who presented for the evaluation of sleep initiation insomnia, while 45% of the patients met clinical diagnostic criteria for RLS and 38% were treated with antidepressants, there were no statistical associations between RLS and antidepressant use or the use of any specific class of antidepressants.(34) However, only two patients in this study were on mirtazapine. Moreover, another study found that SSRIs improved preexisting RLS in 70% of patients, in contrast with 9% of patients who were free of RLS prior to treatment but developed RLS during treatment.(35) A separate study even argued that neither antidepressants nor neuroleptics but nonopioid analgesics are associated with an increased risk of RLS, but this claim has been rebutted based on methodological limitations.(36, 37)
Data that support an association come from a large European epidemiological study involving 18,980 subjects on the study of the prevalence of RLS that reported an increased risk (OR=3.11) in patients taking SSRIs, but not for TCAs or other types of antidepressants.(38) As mentioned earlier, the French pharmacovigilance database also found RLS to be an adverse reaction to antidepressants (OR=15.9; amitriptyline, escitalopram, mianserin, mirtazapine, and duloxetine). In addition, another systematic review provided strong supporting evidence for drug-induced RLS on similar medications.(11, 16) One prospective study evaluated patients who were treated for the first time with antidepressants and sought to answer the question of whether RLS occurred or preexisting RLS worsened as a result of the medication.(39) Although the medications were limited to second-generation antidepressants (fluoxetine, paroxetine, citalopram, sertraline, escitalopram, venlafaxine, duloxetine, reboxetine, and mirtazapine), 9% of patients reported RLS as a side effect of the administration of these drugs, of which mirtazapine represented the most frequent medication causing RLS (28%). Importantly, the symptoms usually occurred during the initial days of treatment with these medications, and abated in the further course of therapy. The co-medication with benzodiazepines, neuroleptics, and TCAs did not pose additional risk of RLS and no case occurred while using reboxetine in this study.
The issue of whether antidepressants can increase the risk of developing RLS is clinically relevant since many RLS patients suffer from depression, and vice versa RLS also occurs more frequently among patients with depression and anxiety.(40, 41) Therefore, the association, if demonstrated, has important clinical implications for how physicians would consider antidepressants in the general population, among whom RLS is a very common disorder, and how they would modify their treatment recommendations in patients who already suffer from RLS and need to start on antidepressants. Even though the evidence is contradictory, the association between RLS and antidepressants is likely to be substantial with considerable differences observed between various agents. While an average risk of SSRIs and SNRIs to induce RLS is about 5%, the risk is probably highest in mirtazapine (28%) and is negligible in reboxetine and bupropion. No studies have specifically determined the risk of RLS in TCAs. Based on the above information, physicians should differentiate the risk of RLS in individual patients who are to be started on antidepressants. In those who are susceptible to RLS due to coexisting risk factors including old age, female, and the presence of chronic diseases (e.g., musculoskeletal disease, iron deficiency, heart disease, obstructive sleep apnea, peripheral neuropathies, Parkinson’s disease), careful considerations should be made when starting antidepressants in these patients, particularly with mirtazapine and mianserin. Once started, physicians should specifically ask the patient for RLS symptoms at the first visit. If no symptoms have occurred by then, RLS is unlikely to appear in the further course of treatment. If the symptoms are not too pronounced, the medication can be maintained with a chance of abatement of RLS in the short term but a close follow-up is recommended. If RLS symptoms are intolerable, physicians may consider switching to another antidepressant with less of a likelihood of precipitating RLS. The most appropriate options include reboxetine and bupropion. A recent double-blinded, randomized controlled trial of 29 moderate to severe RLS patients suggested that bupropion does not exacerbate RLS symptoms and may be a reasonable choice if an antidepressant is needed in individuals with RLS.(42) If a patient who is already on antidepressants complains of RLS symptoms, physicians should determine if a causal relationship with an antidepressant is based on a close temporal association between the start of the antidepressant and the onset of symptoms.
Since a positive therapeutic response to levodopa or a dopamine agonist is considered as a supportive clinical feature of IRLSSG, it seems logical that patients who take antipsychotics may be at risk of developing RLS. Antipsychotics are a class of drugs used to control symptoms in patients with psychotic disorders, such as schizophrenia and delusional disorders. A key pharmacological property of all antipsychotics is their ability to block dopamine-2 (D2) receptors. Distinguished from typical or conventional antipsychotics, “atypical” antipsychotics may be associated at a clinical level with diminished extrapyramidal syndrome, a reduced capacity to elevate plasma prolactin, and less severe negative symptoms.(43) From a pharmacological perspective, the mechanisms of atypical antipsychotics vary depending on individual agents but generally fall into one of four categories: 1) serotonin dopamine antagonists, 2) D2 antagonist with rapid dissociation, 3) D2 partial agonists, and 4) serotonin partial agonists at 5HT1A receptors. These mechanistic properties may be involved in the pathogenesis of RLS in susceptible individuals.(44)
Similar to antidepressants, the majority of the evidence of antipsychotic-induced RLS is in the form of case reports and very few have reported the association of RLS with typical antipsychotics, including levomepromazine, pimozide, haloperidol, and sulpiride.(45–47) The French pharmacovigilance database identified thioridazine and loxapine as the offending agents.(16) Among those atypical antipsychotics, risperidone, olanzapine, aripiprazole, quetiapine, and clozapine are the frequent suspects.(48–59) Olanzapine, risperidone, and aripiprazole were also identified as the frequent causes of drug-induced RLS in two systematic reviews.(11, 16) The low number of reports of RLS in patients taking typical antipsychotics may be due to several causes, such as low awareness of this association when typical antipsychotics were introduced, and does not necessarily imply that atypical antipsychotics are at greater propensity to induce RLS in comparison with typical antipsychotics. Interpretations of those case reports should be made with caution since patients were often co-medicated with other medications, particularly antidepressants and other antipsychotics, and the possibility of either medication or both as a cause of RLS should be considered. Despite some suggestions on the likelihood of various antipsychotics to induce RLS based on their pharmacological profiles, there is currently no published study that specifically evaluates or compares the propensity of different antipsychotics to induce RLS.
The prevalence of RLS in patients taking antipsychotics is expected to be higher than in the normal population. In a study involving 182 hospitalized schizophrenic patients, the prevalence of RLS was significantly higher in the schizophrenia group (21.4%) than in the control group (9.3%).(60) Moreover, the IRLS score of the patients was higher than that of the normal control subjects, indicating that the RLS symptoms were more severe in the patients’ group. The number and dosages of antipsychotics, and co-medications with antidepressants and anxiolytics, were not related to the severity of RLS. A separate study, which determined the prevalence of primary RLS in 100 patients taking antipsychotics for various disorders, found no differences in the prevalence in comparison with the healthy age- and sex-matched controls and only one patient was diagnosed with RLS. The patient developed symptoms four years after taking perphenazine for depression and the association was presumed to be coincidental.(61) The differences in the reported prevalence among these two studies may be due to different methodologies, study populations, and designs. Interestingly, the most common RLS complaints in schizophrenic patients were restlessness and fidgeting, which was different from the control group who reported paresthesia and tingling.(60) Although restlessness and fidgeting could be symptoms of akathisia, these symptoms were very mild or absent during the daytime, posing the possibility that the characters of RLS in schizophrenic patients may be different than idiopathic patients. In most reports, patients developed RLS within hours to days after starting antipsychotics or an increase in the dosage. Symptoms usually disappeared in a matter of hours to days after stopping the medications.(48–51, 53, 54)
The contribution of genetic susceptibility in antipsychotic-induced RLS has been explored in several candidate genes including G protein β3 subunit, D1-4 receptor, MAOA, and MAOB genes.(44, 62, 63) A probable association has been established between antipsychotic-induced RLS and G protein β3 subunit gene and the interaction between the variable number of tandem repeat polymorphisms in the promoter region of the MAOA gene and the A644 G single nucleotide polymorphism in intron 13 of the MAOB gene. This dynamic has been reported to show a significant influence on the RLS scores of patients with schizophrenia.(44)
When considering dopamine receptor blocking agents (DRBAs) as a probable cause of RLS, it is important to evaluate the possibilities of other drugs in addition to antipsychotics that also block dopamine receptors. Among those, calcium channel blockers (cinnarizine, flunarizine), metoclopramide, prochlorperazine, clebopride, domperidone, and melatonin also share at least in part a DRBA action in addition to specific pharmacologic profiles of the individual agents. Prochlorperazine was mentioned as a possible cause of RLS in the original description of Ekbom.(6) Although there was no specific case report of metoclopramide-induced RLS, metoclopramide was tested as a drug challenge in untreated RLS patients with the aim of understanding the dopaminergic mechanism in RLS. Following a 10-mg infusion of metoclopramide, both sensory and motor symptoms of RLS as monitored by suggested immobilization test (SIT) could not be provoked.(64) Both prochlorperazine and metoclopramide can commonly induce akathisia and the clinical differentiation from RLS can be difficult if the symptoms occur in the evening or at night.(65, 66) Interestingly, there was a recent case of idiopathic RLS that dramatically worsened with the addition of domperidone, which is a peripheral D2 receptor antagonist.(14) This possible link was later supported by a study showing that 48% of PD patients who were taking domperidone had RLS, compared with 21% not taking domperidone (p=0.01).(14) As this study determined the prevalence of RLS in PD patients, a true causation between domperidone and RLS in normal populations needs to be evaluated in future studies. One case report of hemi-RLS was induced by clebopride, which is a substituted benzamide with dopamine antagonistic and prokinetic actions similar to metoclopramide.(67) Since melatonin has been shown to inhibit dopamine release in the striatum and there is a temporal relationship between the onset of melatonin secretion at night and the onset or worsening of RLS symptoms, melatonin may theoretically play a role in the genesis of RLS.(68, 69) One small study evaluated the effects of the administration of exogenous melatonin and, conversely, the suppression of endogenous melatonin secretion by bright light exposure in eight RLS patients.(70) Exogenous melatonin significantly increased RLS symptoms while bright light exposure improved leg discomfort slightly. We are not aware of any reports of cinnarizine- and flunarizine-induced RLS in the medical literature.
The above evidence suggests that the associations between RLS and these medications are limited to theoretical possibilities, supported by a few case reports or specific experimental situations. Therefore, a definite conclusion cannot be made as to whether RLS can be a consequence of the use of nonantipsychotic DRBAs.
Evidence on antiepileptic-induced RLS in the medical literature is limited to a few case reports. The first report was from two epileptic patients who developed RLS while taking methsuximide and phenytoin.(71) While the diagnosis of RLS in these two patients was based only on clinical descriptions of creeping sensations in the legs in association with insomnia, the symptoms abated once medications were discontinued. Subsequently, there were reports of RLS as a consequence of topiramate and zonisamide use and the symptoms rapidly resolved as soon as the treatments were suspended.(72–75) The possibility that both topiramate and zonisamide may induce RLS is possibly attributable to their multiple mechanisms of action, including blockage of sodium and T-type calcium channels, inhibition of carbonic anhydrase, inhibition of glutamate release, and dopaminergic activity. Specifically, zonisamide is considered to have biphasic effects on the dopaminergic system in which therapeutic doses have been found to increase intracellular and extracellular dopamine in the rat striatum, while supratherapeutic doses reduce intracellular dopamine.(76) Topiramate has recently been shown to reduce levodopa-induced dyskinesias in MPTP-lesioned primates and modulate cortico-mesolimbic dopamine function.(77, 78) It is not known if these observations are relevant to the occurrence of RLS.
In contrast to a possibility of antiepileptic-induced RLS in very few reported patients, certain antiepileptic drugs have been shown to be efficacious in the treatment of RLS, particularly pregabalin, gabapentin, gabapentin enacarbil, oxcarbazepine, lamotrigine, and levetiracetam. Recent European guidelines on the management of RLS have recommended the use of pregabalin and gabapentin enacarbil as an effective treatment of primary RLS.(79)
While opioids and opioid receptor agonists, including tramadol, methadone, oxycodone-naloxone, and intrathecal morphine, may be considered as alternative medications in the treatment of RLS following unsuccessful treatment with dopaminergic medications, a number of recent reports have highlighted the possibility of tramadol inducing or exacerbating RLS.(11, 16, 80, 81) The French pharmacovigilance database identified three cases of RLS as being attributable to an adverse reaction to tramadol (OR=13.05).(16) Similarly, a recent systematic review with qualitative scoring and comparative analysis suggested that tramadol was among the potential agents that could induce or exacerbate RLS.(11) However, details were unavailable in these two studies as to whether the identified cases were patients with de novo RLS or with augmentation. Two separate case reports suggested the possibility that long-term use of tramadol may augment RLS symptoms.(82, 83)
The combination of dopaminergic agents and an opioid is frequently used in clinical practice, particularly in severe cases, but this approach has never been investigated with respect to efficacy or adverse events. This seems to be a reasonable option for short-term use in severe intractable RLS patients who do not respond to dopaminergic treatment, but physicians should be aware of the possibility of at least augmentation with tramadol. Two recent guidelines have not supported the use of opioids in the treatment of RLS due to insufficient evidence.(79, 84)
There were three case reports of lithium-induced RLS in which the possibility of co-medication with levomepromazine was not excluded in one of the two cases.(46, 85, 86) Although the mechanism of lithium-induced RLS is unclear, the possibility of lithium enhancing noradrenergic functions while reducing dopamine synthesis has been suggested.(85) RLS has been included as a potential side effect among different types of movement disorders in association with interferon-α by exerting complex interactions between dopaminergic transmission and opioid receptors.(87) Methoxy polyethylene glycol-epoetin beta (CERA), a recombinant erythropoietics-stimulating agent without reported mechanism on dopaminergic activity, was recently claimed to induce RLS in one patient with chronic renal failure.(88) However, the possibility of secondary RLS as a result of renal failure and anemia cannot be entirely excluded in this patient. RLS symptoms may also complicate thyroxine replacement in at-risk hypothyroid patients with low serum ferritin.(89) Interestingly, there has been a single case report of one patient with periodic limb movements in sleep (PLMS) with a relatively low ferritin level who developed severe diurnal RLS symptoms a few months after starting a controlled-release levodopa.(90) The symptoms subsided once levodopa was discontinued. This finding may be considered to be a form of augmented RLS symptoms, previously described in patients with PLMS.(17)
It is likely that there will be an increasing number of reports suggesting the association between RLS and various classes of medications in which some of them may share, at least in part, a dopaminergic modulation. At this point, there is insufficient evidence to determine if these medications (in the miscellaneous section) are able to solely induce RLS as an adverse reaction. More likely, the incidence occurs in patients who are posed to be at risk, i.e., those with anemia, low ferritin, or chronic renal failure. Therefore, physicians should be aware of this potential association and exercise caution when these patients return back to them with suggestive RLS symptoms or new-onset insomnia.
Augmentation is the development of progressively more severe RLS earlier in the day after the administration of dopaminergic medications in the afternoon or evening.(17) It may take the form of earlier onset of symptoms (100% of patients), and increasing severity of preexisting symptoms (96%). In addition, patients may experience a spread of symptoms to different body parts, usually the arms, and a shorter duration of the effect of the medication.(91) Moreover, patients with augmentation frequently have an increase in severity with increased doses of medication, and an improvement following a decrease in medication. This paradoxical response to treatment whereby symptoms improve following discontinuation of medication is useful for distinguishing augmentation from disease progression.(92) In the original description, 82% of RLS patients and 31% of PLMS patients developed augmentation following the treatment with levodopa/carbidopa.(17) Augmentation was severe enough to warrant a change in medication in 50% of RLS patients and 13% of PLMS patients. In a recent US community study, estimated prevalence of augmentation in all patients with RLS treated with a dopaminergic medication was 75%, with an annual rate increase of 8%.(93) Four factors were considered to increase the risk of augmentation: 1) dosage and half-life of medications, 2) duration of treatment, 3) low serum ferritin levels, and 4) positive family history of RLS.(94) Subsequent studies have reported different prevalence of augmentation ranging from 27%–82% of patients on long-term levodopa, 8%–56% for those on pramipexole, 3%–9% for those on cabergoline, 9.7% for those on rotigotine, and 4% for patients using ropinirole on a long-term basis.(92, 95–99) An overall augmentation rate of 48% was observed in a study involving 83 patients on a variety of dopamine agonists, with a mean follow-up of 39 months.(100) With levodopa, augmentation may occur within 2 to 4 months.(94) However, it is difficult to compare augmentation rates across drugs, as they have not been adequately evaluated using standardized tools. Nevertheless, it appears that augmentation is more common with levodopa and shorter-acting dopamine agonists and is progressive.(84) Therefore, long-term trials are needed to estimate its real incidence. A recent prospective study specifically designed to determine the incidence of augmentation found that it starts to emerge with a 3.5% incidence of augmentation for ropinirole.(92) With the exception of a case series on the use of tramadol, no cases of augmentation have been reported using nondopaminergic medications.(82, 83)
A consensus on the standardized operational definition of augmentation has been established. The first NIH-sponsored consensus defined the primary feature of augmentation as a drug-induced shift of symptoms to a period of time two hours earlier than was the typical time of daily onset prior to pharmacological intervention.(1) A more refined European RLS study group–sponsored consensus, the Max Planck Institute (MPI) criteria, indicated that reliable detection of augmentation can be obtained based on a four-hour time advance of symptoms, or a smaller (two to four hours) advance of symptoms expressed along with other required clinical indications, including a shorter latency of symptoms at rest, a spread of symptoms to other body parts in addition to the lower limbs, or a greater intensity of symptoms.(101) In addition, the augmentation severity rating scale was developed to measure the severity of augmentation in clinical trials.(101)
Augmentation must be differentiated from conditions that it may closely resemble, including tolerance, early morning rebound, progression of RLS, fluctuations in disease severity, and neuroleptic-induced akathisia.(91) Tolerance occurs when the effectiveness of a medication decreases over time necessitating an increase in dosage in order to maintain the initial relief of symptoms. Therefore, tolerance symptoms do not appear earlier in the day as observed in augmentation. In addition, symptoms do not become more severe than at baseline. As the name suggests, “rebound” refers to the reappearance of symptoms in the early morning, which corresponds to the time when the half-life of the medication has expired.(102) The natural progression of RLS usually takes years to develop, as opposed to augmentation, where symptoms worsen within months. In contrast to augmentation, symptoms of neuroleptic-induced akathisia lack circadian nature and patients often complain of inner restlessness rather than limb restlessness and paresthesia as observed in RLS.(103) Importantly, patients with neuroleptic-induced akathisia usually have a history of neuroleptic exposure and there is usually no family history.
Several mechanisms explaining augmentation have been proposed, including dopamine hyperstimulation, overstimulation of D1 receptors compared with D2 receptors, brain iron deficiency, a genetic predisposition, and changes in neuroendocrine secretion and the chronobiotic system.(91, 94) The full discussion of these mechanisms is beyond the scope of this chapter. However, it is generally believed that small doses of levodopa over a long period of time may result in an increase in the dopamine concentrations in the CNS; and that in the absence of down-regulation, repeated doses can lead to increased dopamine concentrations over a short time frame.(104)
The risk of augmentation will be minimized by treating any possible factors that may contribute to RLS symptoms, such as low ferritin level, and by maintaining the minimal dose of the dopaminergic medications, with the maximal possible effects. The dose increases should be carefully considered, particularly if they exceed usually accepted levels. They should be limited to a breakthrough of clinically important symptoms that cannot be managed behaviorally and should be balanced against the option of adding an alternative medication.(84) When serum ferritin is < 50 μg/L, an oral iron supplement should be considered. Any drugs that potentially exacerbate RLS symptoms, including antidepressants and antipsychotics, should be avoided. Augmentation only requires treatment when it is clinically significant. In very mild cases, patients may increase their daytime activities to diminish the symptoms.(91, 94) If the patient is bothered by symptoms earlier in the daytime, one strategy is to split the medication dose into two or add an additional very small dose of a dopamine agonist.(91, 94) However, this strategy is unlikely to last that long and patients may later develop even earlier symptoms requiring earlier medications. When the symptoms become severe, the approach is to gradually reduce dopaminergic medications and at the same time, add either an α2δ anticonvulsant (e.g., gabapentin enacarbil, pregabalin, or gabapentin), or an opioid.(79) For further details on the management of augmentation, please consult the RLS treatment algorithm and the recent evidence-based guidelines on the long-term management of RLS.(84, 101)